Hey I'm new here but couldn't help chiming in on this discussion. I've needed a reason to do a simple study on mast thickness and this forum finally gave me that reason!
Based on some of the comments here I ran a study in XFLR with my Inde wingset as a baseline, and some different mast sections through a full operating range of speeds. XFLR5 is what I use to do foil design, it's great, its free, and its results are the best out of the handful of CFD panel codes out there. Results are for the full 3D wingset with viscosity included.
The results have following constants:
Chord length: 110mm
Immersed mast section: 380mm (half of a 75cm mast)
NCRIT 1.5 (turbulent flow)
1100cm2 wing and 200cm2 stabilizer
85kg rider
The section is a NACA 6 series scaled to 11% and 17% which results in a 12.3mm and 19mm thickness at this chord length but the section shape other than thickness is the same.
0 degree yaw (important)
Just now realizing I can't post photos, I'll share some interesting graphs once I can.
At 10knots
NACA 64A-011 (12.4mm thick)
Total drag: 52.5N
Mast % of total: 10.48%
NACA 64A-017 (19mm thick)
Total drag: 53.8N
Mast % of total: 12.64%
Difference: 2.16%
At 20knots
NACA 64A-011 (12.4mm thick)
Total drag: 88N
Mast % of total: 18.2%
NACA 64A-017 (19mm thick)
Total drag: 94N
Mast % of total: 23.4%

Difference: 5.22%